Overview
With the increase in the life expectancy in western countries, the question of appropriate care to be given to elderly people has come more to the fore. Technologies that enable people to live comfortably in their own homes even if they suffer mild disability or dementia are an excellent way of empowering them, and delay the need for them to move into a nursing home. The primary cause of dementia today is Alzheimer's disease, and its most apparent symptom is short-term memory loss. This project aims to develop an aid for sufferers of short term memory loss for completing routine (but complex) at-home tasks. Our prior work with elders in the Nightingale project in Australia on Pervasive Computing Applications for memory sharing has informed this project design.
Intel have developed an application where a user wears a glove with an RFID reader embedded in it, which is able to detect when an RFID tag is close to the palm. This application is deployed in an environment containing numerous artifacts that were tagged using RFID. Whenever the user touches any of these artifacts, the reader registers the presence of the object. They demonstrated that their system could accurately recognize the activities that were being undertaken at any time.
The second student will explore the visual exploration of such a large photo corpus from the point of view of each persons social network. A social network is a representation of the people we know and whom those people subsequently know. Here the research will focus on the visual presentation or a medium sized social network which can be searched, annotated and used as a tool for social reminiscence on a tabletop interface, such as the Diamondtouch. The aim is to use the photographic collection as the interface while the set of follow on photos from the current one is driven by your social network and the meta data associated with each photo (event, location, weather, people etc.)
We propose to extend this work. Firstly our system will attempt to understand the context in which an artifact was last handled by the user. If a user is looking for an artifact (e.g., their car keys), the system will provide cues to to jog the user's memory, e.g., ``you left your keys with your mobile phone''. Secondly, this project will extend the work done by Intel on recognizing routine tasks with an extension that looks at ongoing routines and alerts the user if there is a delay in their completion. This is particularly useful for helping users with short-term memory problems, since they often find it difficult to perform routine tasks if they have too many steps. This project will involve the deployment of an RFID rich environment, the development of software for remembering the position or context of tagged artifacts, and the development of a simple task detector that will remind the user when an incomplete task has been abandoned. The range of tasks will be a small set of simple routine tasks from the kitchen domain, which was used by Intel to good effect.
Relevance of Project to the Host Laboratories:
This project is an excellent complement to ongoing work on smart-homes, context awareness, and sensor networks in the Systems Research Group in UCD. If successful, funding will be sought to continue development and examine the feasibility of producing a commercial product